With the exponentially growing use of geographic information systems (GIS) to store, manipulate and visualize geocoded information, it is increasingly important to understand the particular nature of geographic data and the specialized statistical techniques required for its spatial analysis. Spatial analysis is often broadly defined as a “set of methods useful when the data are spatial”. More specifically, it encompasses a collection of techniques to add value to data contained in a geographic information system. As such, spatial analysis forms an important component of the evolving discipline of Geographic Information Science.
    Aggregate spatial data are characterized by dependence (spatial autocorrelation) and heterogeneity (spatial structure). These spatial effects are important in applied statistical analysis, in that they may invalidate certain standard methodological results, demand adaptations to others, and in some contexts, necessitate the development of a specialized set of techniques. These issues are typically ignored by classical statistics and now are vigorously approached in the separate field of spatial statistics.
    In this paper, some general ideas on fundamental issues are outlined, related to the distinctive characteristics of spatial data analysis, as opposed to data analysis in general. The emphasis is on the relevance for spatial data analysis of the ongoing debate about methodology in the disciplines of statistics and econometrics, and on the role of spatial errors in modeling and analysis. First, some general remarks are formulated on two opposing viewpoints regarding spatial analysis and spatial data: a datadriven approach versus a model-driven approach. This is followed by a review of a number of competing inferential frameworks that can be used as the basis for spatial data analysis. Next, the focus shifts to spatial errors and to the implications of various forms of spatial errors for spatial data analysis. Finally, some concluding remarks are formulated on future research directions in spatial statistics and spatial econometrics.